The parton content of virtual transverse photons \ensuremath{\gamma}(${\mathit{P}}^{2}$) is expressed in terms of perturbative pointlike and nonperturbative hadronic (VMD) components. The resulting LO and NLO parton distributions ${\mathit{f}}^{\ensuremath{\gamma}(\mathit{P}2}$)(x,${\mathit{Q}}^{2}$) are smooth in ${\mathit{P}}^{2}$. They apply uniformly to all ${\mathit{P}}^{2}$\ensuremath{\ge}0 whenever \ensuremath{\gamma}(${\mathit{P}}^{2}$) is probed at a scale ${\mathit{Q}}^{2}$\ensuremath{\gg}${\mathit{P}}^{2}$ where the transverse photons furthermore also dominate physically relevant cross sections. Predictions are given for ${\mathit{F}}_{2}^{\ensuremath{\gamma}(\mathit{P}2}$)(x,${\mathit{Q}}^{2}$) and ${\mathrm{\ensuremath{\mu}}}^{\ensuremath{\gamma}(\mathit{P}2}$)(x,${\mathit{Q}}^{2}$), ${\mathit{g}}^{\ensuremath{\gamma}(\mathit{P}2}$)(x,${\mathit{Q}}^{2}$) relevant for future CERN LEP and present DESY HERA measurements, respectively. Except for certain kinematic regions, these virtual structure functions and parton distributions at ${\mathrm{\ensuremath{\Lambda}}}^{2}$\ensuremath{\ll}${\mathit{P}}^{2}$\ensuremath{\ll}${\mathit{Q}}^{2}$ are not solely described by purely perturbative contributions, in contrast with ``naive'' expectations; these remaining nonperturbative components, although being based on VMD-inspired simplicity, represent the largest uncertainty in our quantitative results and eventually have to be tested by future experiments.
Read full abstract